Adaptive controller design using just-in-time learning algorithm

Abstract

In this paper, an adaptive FID controller design based on just-in-time learning (JITL) algorithm is developed. Inspired by the learning strategy of neural network, a gradient descent algorithm is used to tune the controller parameters on-line based on the information provided by JITL. In addition, a predictive feedforward control scheme is devised to enhance the disturbance rejection capability of the proposed design. The proposed method is illustrated and evaluated by a case study of a polymerization reactor. © 2004 IEEE.